See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/263612306

EFFECTS OF ALKALINE ORGANOSOLV DELIGNIFICATION ON ENZYMATIC CONVERSION OF FROM SUGARCANE PRETREATED BY STEAM EXPLOSION

ARTICLE · JANUARY 2013

CITATIONS READS 2 40

6 AUTHORS, INCLUDING:

Viviane Marcos Nascimento Vinícius Fernandes Nunes da Silva Laboratório Nacional de Ciência e Tecnologi… Cia Suzano de Papel e Celulose

11 PUBLICATIONS 50 CITATIONS 14 PUBLICATIONS 84 CITATIONS

SEE PROFILE SEE PROFILE

George Jackson de Moraes Rocha Adilson Roberto Goncalves Laboratório Nacional de Ciência e Tecnologi… National Council for Scientific and Technolo…

64 PUBLICATIONS 1,115 CITATIONS 85 PUBLICATIONS 1,054 CITATIONS

SEE PROFILE SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Vinícius Fernandes Nunes da Silva letting you access and read them immediately. Retrieved on: 13 February 2016 Revista Investigaciones Aplicadas | ISSN 2011-0413 | Medellín - Colombia Vol. 7, No.1 (2013) Enero – Junio | PP. 1-10 http://revistas.upb.edu.co/index.php/investigacionesaplicadas/index

EFFECTS OF ALKALINE ORGANOSOLV DELIGNIFICATION ON ENZYMATIC CONVERSION OF CELLULOSE FROM SUGARCANE BAGASSE PRETREATED BY STEAM EXPLOSION

Luis R. M. Oliveira*, Viviane M. Nascimento**, Débora Lee S. Corso** , Vinícius F. N. Silva*,**, George J. M. Rocha**†, Adilson R. Gonçalves*

* Departamento de Biotecnologia, Escola de Engenharia de Lorena, Universidade de São Paulo,CEP 12.602-810, Lorena, SP, Brazil. ** Brazilian Bioethanol Science and Technology National Laboratory – CTBE, P.O. Box 6170, CEP 13083-970, Campinas, SP, Brazil.

Recibido 07 Mayo 2013; aceptado 23 Mayo 2013 Disponible en línea: 15 Julio 2013

Abstract: The enzymatic conversion of pretreated and delignificated sugarcane bagasse was evaluated by L18 Taguchi experimental design. The factors selected for the study were: and caustic soda concentration, temperature, reaction time and fiber/liquor ratio. The NaOH concentration was the only significant factor in the delignification, hemicellulose solubilization and cellulose conversion. Recovery of ethanol was influenced by only the fiber/liquor ratio, for the same level. The and hemicellulose removal from bagasse pretreated by steam explosion has reached over 80% and 55%, respectively. Pulps with the lowest lignin and hemicellulose contents had the best results of the cellulose conversion, almost 70%.

Keywords: pretreatment, bagasse sugarcane, organosolv, enzymatic conversion, second-generation ethanol.

† Corresponding author: Tel. (+55) 19 3518 3179, fax (+55) 19 3518 3164. E-mail: [email protected] (George Jackson de Moraes Rocha).

L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv….

1. INTRODUCTION washing, leaving a water-insoluble fraction composed of cellulose, residual hemicelluloses Alternatives to petroleum-derived fuels are being and a chemically modified lignin that can be required in order to reduce the world’s further extracted by alkaline treatments (Ramos, dependence on non-renewable resources. The 2003). most common renewable fuel today is ethanol derived from starch and sucrose, arising on the On the other hand, pretreatment alone is not corn grain and the sugar cane, respectively. The sufficient to extensively remove lignin. The lignin extensive researches to new technologies of also forms a protective physical barrier against ethanol production have been developed in the enzymatic attack. So, alkaline extraction or soda last two decades (Mabee and Saddler, 2010) process is used for its removal, contributing to because of the limitations that will probably occur fast enzymatic saccharification of biomass with a in the near future to supply the raw materials of high sugar yield for both hexoses and pentoses this mills. Therefore, the conversion of biomass (Silva et al., 2011). into ethanol has been considered an attractive pathway for future supplies of liquid fuels. These Doherty et al. (2011) comments that the soda technologies include a low-cost thermochemical process is currently the predominant method for pretreatment, highly effective cellulose and lignin extraction of non-wood material, such as hemicellulose hydrolysis and an efficient and bagasse, wheat straw, hemp, kenaf and sisal. The robust fermentative microorganism. organosolv process, which uses an organic solvent in pulping liquor, offers an alternative to Sugarcane bagasse, a waste in the process of conventional delignification processes with sugar extraction, is an abundant and low-cost significant environment advantage (El-Sakhawy lignocelullosic material on alcohol and sugar et al., 1996). This technology is considered a industry. After processing, approximately 280 kg green process, since it reduces production of of bagasse remains per ton of sugarcane solids, liquids and gaseous residual. The (Carrasco et al., 2010). Sugarcane bagasse as well degradation of carbohydrate is also lower than as other types of plant biomass, is composed by soda process (Doherty et al., 2011). The organic cellulose, hemicelluloses, lignin, and small solvent on organosolv delignification promotes amounts of extractives and mineral salts. The the dissolution and reduces condensation structural components are distributed in a lamellar reactions of lignin (Sarkanen, 1990). structure. In addition, the lignocellulosic biomass has also the strong crystalline arrangement of The present study was conducted to determine the cellulose and the protective effects by lignin and effect of ethanol-soda organosolv delignification hemicelluloses that difficult the access of on the cellulose conversion into fermentable enzymes to the β-14 glycosidic bonds of sugars from pretreated sugarcane bagasse. On the carbohydrates (Carrasco et al., 2010). Thus, section 2.1, the lignocellulosic material was pretreatment and delignification processes are submitted to an industrial steam explosion required to disrupt the cellulose-hemicellulose- pretreatment followed by ethanol-soda lignin complex, and they are important delignification, shown on the section 2.2, and technological steps for the fractionation of enzymatic hydrolysis, on 2.4, being characterized lignocellulosic biomass. on section 2.5. Finally, are shown the results and conclusions of the study. The steam explosion pretreatment has been claimed as one of the most successful options for 2. MATERIALS AND METHODS fractionation of lignocellulosic biomass (Ramos, 2003). That consists in a short cooking at high temperature and pressure in saturated steam. 2.1. Pretreatment of lignocellulosic material Following the cooking, the digester is discharged by a fast decompression. The mechanical effect The sugarcane bagasse used in this study was causes a massive disruption of the fibers (Rosa et provided by the Usina de Açúcar, Álcool e al., 1999) and the high-pressure steam modifies Biodiesel Vale do Rosário from Orlândia/SP— the plant cell wall structure, yielding a dark Brazil. The pretreatment of bagasse was carried brown material from which partially hydrolyzed out in the industrial reactors of the Mill. This 3 hemicelluloses are easily recovered by water- company has three 5 m reactors for steam- rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 2 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv…. explosion pretreatment of bagasse which is used hemicellulose solubilization, ethanol recovery to prepare the animal feed. and enzymatic conversion of cellulose.

The pretreatment was performed under 14.5 – Table 1. L18 Taguchi array used in the 15.5 atm pressure (temperature near 200 ºC) for 7 experimental design of organosolv delignification min. This condition was optimized for production of bagasse of animal feed and it was not possible to test other conditions. The idea is employing the same Factor equipment and conditions and thus the Mill could Trial have two processes from only one material. 1 2 3 4 5 6 7 8 Besides, the distance between reactor and cyclone 1 1 1 1 1 1 1 1 1 is very large which prevented to carry out tests 2 1 1 2 2 2 2 2 2 with lower pressures. 3 1 1 3 3 3 3 3 3 4 1 2 1 1 2 2 3 3 After steam-explosion pretreatment, the bagasse 5 1 2 2 2 3 3 1 1 was continuously washed with distilled water to 6 1 2 3 3 1 1 2 2 remove the hydrolyzed hemicellulose fraction in 7 1 3 1 2 1 3 2 3 order to be used in studies to obtain xylo- 8 1 3 2 3 2 1 3 1 oligosaccharides and xylitol by fermentation 9 1 3 3 1 3 2 1 2 pathway, within the concept of biorefinery. The 10 2 1 1 3 3 2 2 1 washed material was stored in 60 L bags after 11 2 1 2 1 1 3 3 2 sundrying. 12 2 1 3 2 2 1 1 3 13 2 2 1 2 3 1 3 2 2.2. Experimental design of organosolv 14 2 2 2 3 1 2 1 3 delignification 15 2 2 3 1 2 3 2 1 16 2 3 1 3 2 3 1 2 An organosolv process was used to study the 17 2 3 2 1 3 1 2 3 lignin removal of the pretreated bagasse. The 18 2 3 3 2 1 2 3 1 organosolv processes is based on the use of organic solvents that promote the dissolution and reduces condensation reactions of lignin Table 2. Factor levels of L18 Taguchi array used (Hongzhang and Liying, 2007). The ethanol-soda in the experimental design to organosolv organosolv process was chosen for delignification delignification of bagasse of the pretreated bagasse, since the presence of sodium hydroxide improves the delignifying Level ability of ethanol (Muurinen, 2000). Factor 1 2 3

For this study, a L18 Taguchi array was used to Ethanol (%) 40 60 80 evaluate the parameter effects of the organosolv Temperature 100 130 160 delignification. This array was chosen because it (ºC) is able to include several factors in a reduced Time (min) 30 60 90 account of trials. This type of array has 8 NaOH (%) 0.1 0.55 1.0 maximum numbers of factors, one of them with Fiber/Liquor two levels whereas the others with three levels. 1:10 1:7.5 1:5 For this work, only five columns of the array Ratio were used, factors 2, 3, 4, 5 and 6. The factors of experimental design were concentration of The trials of L18 Taguchi array were carried out ethanol and caustic soda, temperature, reaction in a 500 mL stainless steel ampoule heated by time and fiber/liquor ratio and are shown on silicone bath. 30 g of steam-explosion bagasse Table 1. The levels for each factor are shown on were used to all trials of experimental design. Table 2, which were selected according to After reaction time, the ampoule was quenched Hongzhang and Liying (2007). The factors 1, 7 by ice bath. After that, the black liquor was and 8 were used to calculate standard error of the separated by filtration, the solid was washed with experimental design. The response variables of 2 L distilled water at 70ºC and the ethanol in Taguchi array were delignification rate, liquor was evaporated in rotary evaporator. A rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 3 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv…. sample of 5 mL of distillate was added in one 0.05 mol.L-1 sodium citrate buffer pH 4.8 bottle properly weighed, with the aid of a supplemented with 0.02% (w/v) sodium azide to volumetric pipette, to determine the mass of the inhibit microbial contamination, and then mixed corresponding volume. Thus, it was possible to with the at 1:10 (w/v) fiber/liquor ratio. determine the density of the distillate. The experiments were carried out in 125 mL Erlenmeyer flasks containing 30 mL total reaction The ethanol concentration in distillate was volume. The flasks were sealed with a plastic film calculated according to the expression and incubated in a rotary shaker at 100 rpm, 45ºC during 72 h. 1− d % (v / v) =100⋅ distilled , ETHANOL 1− d (1) Pulps with 70% moisture were employed to aid Ethanol the self soaking. After 72 h reaction, the hydrolysates were heated for 5 min in a boiling where ddistilled is the distilled density (g/mL) and water bath to precipitate the protein and prevent dEthanol is the density of pure ethanol at 25ºC further hydrolysis. The fiber residue was (0.791 g/mL), value obtained in the label of separated of hydrolysate by paper filtration and product. washed with 15 mL of distilled water. Sugars contained in filtrate were quantitatively The percentage of solvent recovery was determined by high performance liquid calculated from the ethanol concentration and the chromatography (HPLC). volume of liquor. The analysis of the results of experimental design was performed using the The cellulose conversion ratio was calculated software STATISTICA version 7.0 of the according to the expression StatSoft Company. mglu cos e × f h CC = ×100 , (2) m × y 2.3. Enzymes initial i

The enzymes used were commercial cellulase where CC is the enzymatic cellulose conversion, concentrates, the Celluclast 1.5 L, produced by mglucose is the glucose mass in the hydrolysate (w), Trichoderma reesei, supplemented with Novozym minitial is the dry mass of lignocellulosic material, 188, the β-glucosidase, produced by Aspergillus before enzymatic hydrolysis (w), yi is the niger, kindly supplied by Novozymes Latin cellulose content in the lignocellulosic material America Ltd. and fh is the hydrolysis factor of cellulose, correspondent to 0.9. activity of the commercial cellulase concentrate was measured according to Mandels 2.5 Chemical characterization of et al. (1976), and expressed in filter paper units lignocellulosic materials (FPU). The formed reducing sugar was estimated by dinitrosalicylic acid (Miller, 1956). One unit The chemical composition: cellulose, of FPU is defined as the amount of enzyme hemicelluloses and lignin, of the raw material, required to liberate 1 µmol of glucose from nº 1 pretreated bagasse and cellulose , was Whatmann filter paper per minute at 50ºC. determined according to Gouveia et al. (2009). Activity of the β-glucosidase was measured Carbohydrates and organic acids were determined according to Mongkolthanaruk and Dharmsthiti by HPLC in a Shimadzu LC-10AD (2002). chromatograph equipped with a Shimadzu RID- 6A refractive index detector and a Aminex HPX- 2.4. Enzymatic hydrolysis 87H, 300 9 7.8 mm, Bio-Rad Laboratories Ltd, column. The samples were diluted with deionized Only one condition of enzymatic hydrolysis was water, filtered through Sep-Pak C18 filters, used in order to evaluate only the effect of Millipore, and thus injected in the chromatograph organosolv delignification parameters on the under these conditions: 45ºC column temperature, cellulose conversion of delignified bagasse. 0.005 mol.L-1 sulfuric acid as mobile phase at 0.6 mL.min-1 flow rate. The cellulose, 15 FPU/g substrate, and β- glucosidase, 10 IU/g substrate, were added to rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 4 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv….

3. RESULTS AND DISCUSSION the prediction condition. None of the trials in the design of experiments have this combination of The chemical characterization of the used bagasse factor levels. is shown on Table 3. The data are similar to those Average Eta by Factor Levels reported by the literature (Silva et al., 2011; Mean=61.1594 Sigma=13.1337 MS Error=13.5572 df=2 (Dashed line indicates ±2*Standard Error) Rocha et al., 2012). 75

Table 3. Chemical composition in percentage, of 70 the in natura bagasse and the pretreated bagasse. 65

60 Component In natura Pretreated bagasse bagasse 55 Delignification (%) Delignification Yield 68,0% 50

Cellulose 44,0 ± 1,0 51,7 ± 0,6 45 Hemicellulose 25,8 ± 0,8 8,9 ± 0,1 40 Lignin 28,2 ± 0,8 34,3 ± 0,3 1 2 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 Ash 1,4 ± 0,2 5,5 ± 0,2 Factor 1 Ethanol Temperature Time NaOH F:L Ratio Factor 7 Factor 8 Total 99,2 ± 0,8 100,3 ± 0,4 Fig. 1. Graphic generated from the analysis of The analysis of Table 3 revealed that most of the main effects on the response variable hemicellulose is solubilized once the delignification rate. concentration was reduced to 8.9% after the pre- The analysis of variance (ANOVA) for the treatment, while in sugarcane bagasse represented delignification rate pointed out only the caustic approximately 26% of the composition of the soda concentration as a significant factor in material. These data show, beyond the chemical organosolv delignification, NaOH p-value < 0.05 composition, that pre-treatment with steam or F value greater than Ftabulated, indicated on explosion has the characteristic of preferentially Table 5. removing the hemicellulose in the cellulose. Other factors have p-values higher than 0.05 and A L18 Taguchi array was used to evaluate F value less than Ftabulated, on Table 5 Ftab is 4.74 influence of ethanol and caustic soda for a confidence level of 95%. Their factor square concentration, reaction time, temperature and sums were also similar to the empty column fiber/liquor ratio on organosolv delignification of square sums, the errors. This indicates these steam-explosion pretreated bagasse. The response factors have no influence on response variable, variables of Taguchi array were: delignification the delignification rate, for a confidence level of rate, hemicellulose solubilization, ethanol 95%, on Fig. 1 is also possible to verify this recovery and enzymatic conversion of cellulose. observation. Only the NaOH concentration factor The results are depicted in the Table 4. caused a change in the dependence variable Delignification rates for pretreated bagasse higher than limits of red dashed lines; this line obtained in the design of experiments ranged represents the deviation of two standard errors from 40 to 82%. The best result was trial number about the mean value. The variation of other 10, since it caused 82% of delignification of factors was less than limits of the red dashed lignocellulosic material, shown on Table 3. lines. Sarkanen (1990), in his organosolv pulping However, the prediction condition suggested by study comments that the solvent would not be the analysis of main effects that maximizes the likely to change the reactivity of lignin with delignification rate is different from trial number respect to NaOH significantly. The beneficial 10. This analysis indicates the condition: ethanol effect of solvent may be ascribed to either concentration and fiber/liquor ratio at level 2, improved lignin solubility or its reduced tendency 60% (v/v) and 1:7.5 (w/v), respectively, NaOH towards condensation. The results on Fig. 1 show concentration and temperature at level 3, 1% that ethanol solvent has an influence on the (w/v) and 160 °C, respectively, and reaction time delignification rate. The increase of solvent at both 1 and level 2, since these levels have the concentration from 40 to 60% caused an increase same influence on organosolv delignification, of delignification in the lignocellulosic material, Fig. 1. The shortest time was chosen, 30 min, as but its influence was not considered statistically significant for a confidence level of 95%. rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 5 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv….

Table 4. Results of response variables of the L18 Taguchi array used in the design of experiments to organosolv delignification of the bagasse pretreated by steam explosion

Ethanol Temperature Time NaOH F:L Ratio Delignif. Hemicellu. Ethanol Cellulose Trial Rate Solubiliz. Recovery Conversion % (v/v) (ºC) (min) % (w/v) (w:v) (%) (%) (%) (%) 1 40 100 30 0.1 1:10 46.4 23.3 28.2 45.9 ± 0.9 2 40 130 60 0.55 1:7.5 73.3 46.2 27.5 62 ± 1 3 40 160 90 1.0 1:5 71.8 56.1 31.2 67.1 ± 0.1 4 60 100 30 0.55 1:7.5 66.9 41.6 55.3 62.9 ± 0.7 5 60 130 60 1.0 1:5 68.3 41.1 21.7 57 ± 2 6 60 160 90 0.1 1:10 52.7 19.3 63.7 42 ± 2 7 80 100 60 0.1 1:5 40.9 14.3 21.7 38 ± 3 8 80 130 90 0.55 1:10 58.2 43.5 66.8 55 ± 1 9 80 160 30 1.0 1:7.5 68.0 55.8 59.2 57.6 ± 0.3 10 40 100 90 1.0 1:7.5 82.0 50.8 60.3 69.9 ± 0.4 11 40 130 30 0.1 1:5 41.2 15.1 24.9 44.8 ± 0.7 12 40 160 60 0.55 1:10 62.2 39.1 42.9 54 ± 1 13 60 100 60 1.0 1:10 74.7 55.9 63.4 63.8 ± 0.8 14 60 130 90 0.1 1:7.5 52.7 25.8 38.0 50 ± 1 15 60 160 30 0.55 1:5 72.4 46.2 35.7 70 ± 1 16 80 100 90 0.55 1:5 44.3 24.3 25.2 44 ± 2 17 80 130 30 1.0 1:10 75.0 56.3 69.4 68 ± 2 18 80 160 60 0.1 1:7.5 50.8 24.9 49.8 33.9 ± 0.9

Table 5. ANOVA data of the factors for hemicellulose solubilization, as shown in Fig. 2 delignification rate from experimental design. and Table 6. (SS stands for sum of square. SM symbolizes Average Eta by Factor Levels mean of square) Mean=37.7889 Sigma=15.0169 MS Error=95.8172 df=2 (Dashed line indicates ±2*Standard Error) 55

Degree p- 50 Effect SS SM F free value 45

Factor 1 3.47 1 - - - 40

Ethanol 244.96 2 122.48 2.50 0.151 35

Temperature 47.11 2 23.55 0.48 0.637 30 Time 9.55 2 4.77 0.10 0.908 25

NaOH 2026.15 2 1013.08 20.69 0.001 Hemicellulose Solubilization(%) 20 F:L ratio 261.98 2 130.99 2.68 0.137 15 Factor 7 258.75 2 - - - 1 2 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 Factor 1 Ethanol Temperature Time NaOH F:L Ratio Factor 7 Factor 8 Factor 8 53.32 2 - - - Residual 342.66 7 48.95 - - Fig. 2. Graphic generated from the analysis of Total 2932.41 18 - - - main effects on the response variable hemicellulose solubilization.

Another component of lignocellulosic that could P-value of this factor was less than 0.05 and its F be affected by organosolv delignification is value was greater than Ftabulated, which on Table 6 hemicellulose. The solubilization of this is 4.74 for a confidence level of 95%. The other component in the design of experiment ranged factors had the F value lower than Ftablulated 14.3 to 56.3%. The best results, shown on Table indicating to have no influence on the response 4, were obtained from the experiments 3, 9, 13 variable, the hemicellulose solubilization, for a and 17, causing one solubilization of above 55%. confidence level of 95%. The square sums for The ANOVA and main effects indicated only the these factors were lower or similar than the error NaOH concentration as significant factor for the factors, except to the fiber:liquor ratio factor. The best condition for the NaOH concentration factor rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 6 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv…. was 1%, level 3 on Fig. 2. The average of the process. At higher fiber/liquor ratios, solvent hemicellulose solubilization at this condition on loss is reduced due to the great liquor volume Fig. 2 was 53%, close to the values obtained in available in this process. the trials 3, 9, 13, and 17. In our study, the fiber/liquor ratio of 1:10 was Table 6. ANOVA data of the factors for chosen as the largest in the design of experiments hemicellulose solubilization from experimental in order to reduce the solvent consumption. At design. (SS stands for sum of square. SM this condition, the solvent loss was 2.6 L per symbolizes mean of square) kilogram of lignocellulosic, taking into account a solvent recovery average of 56% and ethanol Degree p- concentration average of 60%, while for Effect SS SM F free value Hongzhang and Lying (2007) the ethanol loss Factor 1 0.27 1 - - - was 3.0 L per kilogram of lignocellulosic, which Ethanol 12.41 2 6.20 0.15 0.8657 used a 1:50 fiber/liquor and that had 40% ethanol Temperature 81.99 2 41.00 0.97 0.4239 concentration and 85% solvent recovery. Time 37.25 2 18.62 0.44 0.6596 Therefore, since using a fiber/liquor lower, our study has an absolute loss 15% lower than those NaOH 3183.27 2 1591.64 37.77 0.0001 the authors. F:L ratio 223.74 2 111.88 2.65 0.1387 Factor 7 76.07 2 - - - Table 7. ANOVA data of the factors for ethanol Factor 8 27.00 2 - - - recovery from experimental design. SS stands for Residual 294.98 7 42.14 - - sum of square. SM symbolizes mean of square Total 3833.64 18 - - -

Degree p- Effect SS SM F One advantage of organosolv delignification is free value the solvent recovery used in the process, and the ability to produce substantial amounts of useful Factor 1 65.59 1 - - - by-products, like lignin, sugars, furfural and Ethanol 561.89 2 280.94 2.33 0.167 , as well as pulp (Hongzhang and Temperature 112.25 2 56.12 0.47 0.646 Liying, 2007). The ethanol recovery in the Time 313.56 2 156.78 1.30 0.331 experiment design ranged from 22% to 70%. The NaOH 537.62 2 268.81 2.23 0.178 ANOVA and main effects only pointed out the 11.3 F:L ratio 2723.06 2 1361.53 0.006 fiber/liquor ratio as a significant factor for ethanol 0 recovery. Its P-value was less than 0.05 and F Factor 7 554.26 2 - - - value was greater than Ftabulated, which on Table 7 Factor 8 2.64 2 - - - is 4.74 for a confidence level of 95%. Error 843.26 7 120.47 Total 5091.64 18 - The other factors had the square sums similar than the empty columns, errors, indicating the Average Eta by Factor Levels Mean=43.6056 Sigma=17.3024 MS Error=110.761 df=2 other factors have no influence on response (Dashed line indicates ±2*Standard Error) variable, the ethanol recovery, for a confidence 60 level of 95%. The best condition for the 55 fiber/liquor ratio factor was 1:10, level 1 on Fig. 50

3, which had a solvent recovery average of 56%, 45 value considered insufficient to make feasible a 40 solvent recovery unit. 35 EthanolRecovery (%) The higher fiber/liquor ratio used in our 30 experiments was only 1:10, a ratio lower than the 25 values employed in the literature. Hongzhang and 20 Liying (2007) evaluated the clean fractionation of 1 2 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 Factor 1 Ethanol Temperature Time NaOH F:L Ratio Factor 7 Factor 8 wheat straw using the steam explosion pretreatment and ethanol-water delignification. In Fig. 3. Graphic generated from the analysis of this study, the authors used a fiber/liquor ratio of main effects on the response variable ethanol 1:50 (w/v), giving an ethanol recovery of 85% in recovery. rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 7 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv….

The amount of hemicellulose and lignin in the had p-value higher than 0.05 and F value smaller biomass is the main influence factors of than Ftabulated and their square sums were similar enzymatic hydrolysis yield (Öhgren et al., 2007). than the square sums of the empty columns, the This information can be confirmed by the linear errors. This indicates that these factors have no correlation among cellulose conversion and influence on response variable, the cellulose solubilization of lignin and hemicelluloses, shown conversion, for a confidence level of 95%. El- on Fig. 4. The higher cellulose conversions were Sakhawy et al. (1996), in his study of alcoholic- obtained to the pulps with more solubilization of alkaline pulping for wheat straw, also noted that lignin and hemicellulose, trials 3, 10, 13 and 17 concentration of alkali had a major influence on on Table 4. Hemicellulose forms a physical delignification of lignocellulosic material. barrier around the cellulose hindering enzyme Average Eta by Factor Levels access to cellulose. Lignin also forms a protective Mean=54.5804 Sigma=11.3154 MS Error=38.9447 df=2 (Dashed line indicates ±2*Standard Error) physical barrier to enzymatic attack and it causes 64 inhibition of hydrolysis by non-specific 62 60 hydrophobic binding of lignin to the cellulose 58 binding domain of the enzymes (Berlin at al., 56 2006; Gray et al., 2006; Öhgren et al., 2007). The 54 correlations obtained in this work could be used 52 50 to estimate the cellulose conversion of sugarcane 48 bagasse pretreated by steam explosion and CelluloseConversion (%) 46 delignified by alkaline organosolv, provided that 44 42 same conditions of pretreatment, organosolv 40 delignification and enzymatic hydrolysis of 1 2 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 Factor 1 Ethanol Temperature Time NaOH F:L Ratio Factor 7 Factor 8 cellulose are used. Fig. 5. Graphic generated from the analysis of 80% Lignin Hemicellulose main effects on the response variable cellulose

70% conversion. y = 0.6707x + 0.2925 R² = 0.7884 60% Table 8. ANOVA data of the factors for cellulose conversion from experimental design. (SS stands 50% y = 0.7876x + 0.0641 for sum of square and SM symbolizes mean of R² = 0.8356 40% square) Cellulose Conversion (%)

30% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Degree p- Effect SS SM F free value Solubilization(%) Factor 1 4.98 1 - - - Fig. 4. Linear correlation among cellulose Ethanol 259.13 2 129.57 3.32 0.097 conversion and solubilization of lignin and Temperature 13.66 2 6.83 0.18 0.843 hemicellulose from the design of experiments. Time 139.73 2 69.86 1.79 0.235 NaOH 1470.97 2 735.49 18.85 0.002 Data from experiment design show the removal F:L ratio 20.01 2 10.01 0.26 0.781 of lignin and hemicellulose was affect by only the Factor 7 142.34 2 - - - NaOH concentration in the experiment design. Factor 8 47.93 2 - - - Therefore, it is expected that cellulose conversion will also be influenced by the same factor. The Residual 273.13 7 39.02 analysis of main effects indeed shows this effect Total 2176.63 18 - - - on the cellulose conversion. Only the NaOH concentration caused a variation of the results The prediction condition of organosolv higher than limits of red dashed lines, this line delignification for efficiently conversion of represents the deviation of two standard errors cellulose from lignocellulosic material was: about the mean value on Fig. 5. The ANOVA ethanol concentration, temperature and also pointed out this behavior. Only the NaOH fiber:liquor ratio at level 2, 60% (v/v), 130 ºC and concentration has a p-value less than 0.05 and F 1:7.5 (w/v), respectively, reaction time at first level - 30 min - and NaOH concentration at level value greater than the Ftabulated, which is 4.74 for a confidence level of 95% on Table 8. Other factors 3 -1% (m/v). None of 18 points tested in the L18 rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 8 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv….

Taguchi array contained these conditions. As a process to fractionate the bagasse pretreated by result, a trial of organosolv delignification under steam explosion. NaOH concentration was the prediction conditions was carried out in order to only significant factor of experiment design. The validate the model used in the analysis of the removal of lignin and hemicellulose was essential results from design of experiments. The results of to increase cellulose digestibility of the bagasse this test were displayed in Table 9. pretreated by steam explosion at 200°C/7min. The pulps with higher solubilization of lignin and Table 9. Data of variable responses for the hemicellulose had the best results of cellulose prediction condition test of the organosolv conversion, almost 70%. delignification. Ethanol concentration: 60% (v/v), temperature: 130 °C, time: 30 min, NaOH concentration: 1% (w/v) and fiber/liquor ratio: ACKNOWLEDGEMENT 1:7.5 (w:v) This work was supported by Fundacão de Amparo à Pesquisa do Estado de São Paulo Response Variable Value (FAPESP), Coordenação de Aperfeiçoamento de Delignification rate 74.1 % Pessoal de Nível Superior (CAPES) and Conselho Nacional de Desenvolvimento Hemicellulose solubilisation 56.1 % Científico e Tecnológico (CNPq). The authors are Ethanol recovery 45.0 % grateful to Novozymes Latin America Ltd by Cellulose conversion 62 ± 1 % supplying of the enzymes and to Usina de Açúcar, Álcool e Biodiesel Vale do Rosário by providing the steam-pretreated sugarcane Value of delignification rate obtained for test trial bagasse. on Table 9 was 9.8% lower than trial number 10, which had the best results of the experiment REFERENCES design – Table 4. The reduction of delignification rate caused a decline in the cellulose conversion, - Berlin, A., Balakshin, M., Gilkes, N., Kadla, J., 11,3%, despite the hemicellulose removal has Maximenko, V., Kubo, S. and Saddler, J. increased in the trial test - 10.2%. The inhibitory (2006). Inhibition of cellulase, xylanase and β- effect of lignin on cellulases was more significant glucosidase activities by softwood lignin with respect to increase the cellulose accessibility preparations. J. Biotechnol. 125: 198–209. by reduction of hemicellulose content. Carrasco, C., Baudel, H.M., Sendelius, J., Modig, T., Roslander, C., Galbe, M., Hahn-Hägerdal, The cellulose conversion of 62% obtained for the B., Zacchi, G. and Lidén, G. (2010). SO2- test trial is lower than the expected value from the catalyzed steam pretreatment and fermentation experiment design analysis. We conclude that the of enzymatically hydrolyzed sugarcane array model used in the design of experiments is bagasse. Enzyme Microb. Tech. 46: 64–73. not suitable to depict the behavior of the response Doherty, W.O.S., Mousavioun, P. and Fellows, variable in relation to the organosolv C.M. (2011). Value-adding to cellulosic delignification factors. The interactions among ethanol: Lignin polymers. Ind. Crops Prod. 33: factors could be occurred in the experimental 259–276. design and the model used to maximize the El-Sakhawy, M., Lönnberg, B., Fahmy, Y. and conversion of cellulose could be deficient, since Ibrahim, A. A. (1996). Organosolv pulping. 3. the L18 Taguchi arrays did not allow to evaluate Ethanol pulping of wheat straw. Cell. Chem. effects of the factor interactions on the response Technol. 30: 161-174. variables (Ross, 1991). New experiments to Gouveia, E.R., Nascimneto, R.T., Souto-Maior, confirm the possibility of the interaction among A.M. and Rocha, G.J.M. (2009). Validação de factors will be carried out in the future works. metodologia para a caracterização química de bagaço de cana-de-açúcar. Química Nova, 32 4. CONCLUSION (6): 1500-1503. Gray, K.A., Zhao, L. and Emptage, M. (2006). The results from experiment design for steam- Bioethanol. Curr. Opin. Chem. Biol. 10: 141– explosion pretreated bagasse showed that the 146. ethanol-soda delignification is an effective rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 9 L. Oliveira, V. Nascimento, D. Corso, V. Silva, G. Rocha, A. Gonçalves | Effects Of Alkaline Organosolv….

Hongzhang, C. and Liying, L. (2007). Unpolluted obtained by enzymatic saccharification of fractionation of wheat straw by steam sugarcane bagasse pretreated by hydrothermal explosion and ethanol extraction. Bioresource processing. J. Ind. Microbiol. Biot. 38 (7): 809- Technol. 98: 666-676. 817. Mabee, W.E. and Saddler, J.N. (2010). Bioethanol from lignocellulosics: Status and ABOUT THE AUTHORS perspectives in Canada. Bioresource Technol. 101: 4806–4813. Luis Ricardo Martins Oliveira Mandels, M., Andreott, R. and Roche, C. (1976). Measurement of saccharifying cellulose. Doctor from the University of São Paulo in Biotechnology and Bioengineering Symposium Industrial Biotechnology. Area of investigative 6: 2-34. interest: Conversion of plant biomass. Miller, G.L. (1956). Use of dinitrosalicylic acid Viviane Marcos Nascimento reagent for determination of reducing sugar. Anal. Chem. 31: 426-428. Chemistry in National Laboratory of Science and Mongkolthanaruk, W. and Dharmsthiti, S. (2002). Technology of Bioethanol (CTBE) at the National Biodegradation of Lipid-rich wastewater by Center for Research in Energy and Materials mixed bacterial consortium. Dech. Monog. 50: (CNPEM) and PhD in Chemical Engineering 101-105. from the Universidade Estadual de Campinas Muurinen, E. (2000). Organosolv Pulping – A (Unicamp). review and distillation study related to peroxyacid pulping. En: Academic Débora Lee Simões Corso Dissertation, Faculty of Technology, University of Oulu, Linnanmaa. Graduate Student in Chemical Technology from Öhgren, K., Bura, R., Saddler, J. and Zacchi, G., the Universidade Estadual Paulista (UNESP), and (2007). Effect of hemicellulose and lignin an intern at the National Laboratory of Science removal on enzymatic hydrolysis of steam and Technology of Bioethanol (CTBE) at the pretreated corn stover. Bioresource Technol. 98 National Center for Research in Energy and (13): 2503–2510. Materials (CNPEM). Ramos, L.P. (2003). The chemistry involved in the steam treatment of lignocellulosic Vinícius Fernandes Nunes da Silva materials. Química Nova 26(6): 863-871. Doctorate in Biochemical Engineer from the Rocha, G.J.M. Gonçalves, A.R. Oliveira, B.R. School of Engineering of Lorena (EEL), Oliveira, E.G. Rossell, C.E.V. (2012). Steam University of São Paulo, in collaboration with explosion pretreatment reproduction and National Laboratory of Science and Technology alkaline delignification reactions performed on of Bioethanol (CTBE) at the National Center for a pilot scale with sugarcane bagasse for Research in Energy and Materials (CNPEM). bioethanol production. Industrial Crops and Products. 35: 274-279. George Jackson Moraes Rocha Rosa, M., Rocha, G. and Curvelo, A. (1999). Organosolv delignification of steam exploded Senior Researcher in Industrial Program (PIN) at wheat straw. RECEN - Revista Ciências the National Laboratory of Science and Exatas e Naturais, Guarapuava, 1, 49–55. Technology of Bioethanol (CTBE) in the Ross, P.J. (1991). Aplicações das Técnicas National Center for Research in Energy and Taguchi na Engenharia da Qualidade. Materials (CNPEM). (McGraw-Hill Ed), São Paulo. Sarkanen, K.V. (1990). Chemistry of solvent Adilson Roberto Gonçalves pulping. Tappi J. 73: 215-219. Professor at the School of Engineering of Lorena Silva, V.F.N., Arruda, P.V., Felipe, M.G.A., (EEL), University of São Paulo. Gonçalves, A.R. and Rocha, G.J.M. (2011). Fermentation of cellulosic hydrolysates

rev.investig.apl | Vol. 7, No.1 (2013) Enero – Junio | p. 10